Arguments
The strong economics of wind energy
Posted on 28 December 2015 by John Abraham
As a follow-up to a recent article I posted on renewable energy, this article discusses the economics of wind in both the developed and developing worlds compared to other renewable energy sources. At the recent climate conference in Paris, 70 countries highlighted wind as a major component for their emissions-reduction schemes.
I spoke with Giles Dickson who is CEO of the European Wind Energy Association(EWEA). I asked him economic questions related to the wind industry and I also asked him to look into his crystal ball and describe the future of wind. Mr. Dickson is in a great position to answer these questions because his organization includes 600 members who represent wind industry manufacturers, operators, and companies comprising the full wind-energy supply chain.
First, I asked why companies were investing in wind. His response was clear: wind is competitive economically. He told me about the SolutionWind campaign which is a platform that gives industry leaders like Unilever, BNP, Aveda, IKEA, LEGO, Google, Microsoft, SAP, and others the chance to tell their customers and the general public why they have chosen wind. SolutionWind includes interviews with these leaders (and case studies soon to be published) wherein the case is made that using wind energy adds value to these companies.
Companies want to reduce their emissions and they want access to reliable, inexpensive power. Companies want to know how to achieve these two goals in a way that is quick and efficient. For many of them, wind is the answer. It’s inexpensive and emissions-free (aside from initial manufacturing and installation and service) and it gives the companies control over their energy supply.
Globally, the average cost of wind is $83 per megawatt-hour. This is the levelized cost of electrical delivery. How does it compare to other energy sources? Well the averages for coal and gas are $84 and $98, respectively. In the USA, gas is slightly cheaper than wind but this is the only large economy where that is the case. As a comparison, solar photovoltaic energy averages $122 globally for each MW-hour.
Levelized cost to produce electricity outside of Europe, from Bloomberg New Energy Finance.
There are some additional system integration costs and market balancing costs that vary geographically. The cost of balancing out the variable wind power is usually paid by the wind-power producers. And of course there are the costs of reinforcing the grid, such as building transmission lines to wind farms.
Put together all these additional costs might typically amount to 30% of the total costs. However, they are reduced with a properly functioning electrical market, which balances out variable power from power sources over a large geographical region.
In Europe, for instance, while electricity is traded across borders, there is no single market. That makes the system less efficient. It’s important to include large regions in a single market because a quiet day in Spain may be balanced by wind gusts in Germany. And the same is true for solar. Cloudy days in one region are often balanced by sun in others.
One of the things Mr. Dickson was most excited about was the continued decrease in cost and increase in capacity. The economics of wind are going to continue to get better and installation will accelerate, particularly in the developing world.
Since my research is on small-scale off-grid wind power, I asked him about that too. He told me that small-scale wind is not a big presence in Europe; the vast majority of their wind is grid-connected farms. There are small-scale installations that communities have established for their own local power however micro-generation in Europe is much more likely to be solar rather than wind.
In the developing world, it is a different story. Small-scale off-grid systems are seen as a quick way to extend power to people who do not have access to the grid. According to Mr. Dickson, if your goal is to get people access to quick electricity, off-grid systems have real merit. However if your goal is to provide a basis for long-term energy supply and industrial development, then large-scale systems will be required.
The renewable energy sector also supplies many high-paid and high-skilled jobs. In Europe, for instance, the coal industry employs 240,000 workers whereas the wind industry exceeds that at 262,000. So, people can see careers in this industry and young engineers and scientists are passionate about working in the field. I see this with my own students in the USA.
What does the term "levelized", in the context of this article, mean? Thanks
LOC -> The per kilowatt hour cost of building and operating a power plant over an assumed financial life and duty cycle.
In general, you cannot compare directly non-dispatchable power sources like wind and solar with dispatchable sources like coal, gas, and nuclear.
If the wind farms had sufficient backup power storage on site to supply power for several hours (or perhaps several days) of low wind, then the levelized cost of the combined wind turbine + power storage unit could indeed be compared directly to other dispatchable power sources.
--------------------------
The article says "The cost of balancing out the variable wind power is usually paid by the wind-power producer". Does this mean that the wind-power producer pays to keep a coal or gas plant on hot standby? Please explain what sort of balancing out is paid for by the wind-power producer during low wind days.
Incredible! You are doing it again. Uncritically parroting whatever lobbyists tell you. Your figure on costs rely on gray literature which in fact seems to be unavailble outside press releases and at least in the case of nuclear costs widely inconsistent with serious sources. This type of "reporting" just undermines the good work you do.
The table is wildly underestimating the cost of nuclear if it doesn't include the costs of decommissioning and 'storage' of waste for hundreds of years.
Who pays for that?
jpjmarti @4... The EIA puts LCOE for onshore wind at $73/kWh in the US [link], so the numbers are clearly not out of line, and even are likely understated in this article.
Charlie A - Several points:
On a one-to-one generator basis, no, renewables and fossil fuels are not directly comparable. On a system-to-system basis, they are quite comparable, and renewable baseload power is indeed less expensive for economies on a whole.
Charlie A @3, the demand for 'dispatchable' power is overstated. Currently, with dispatchable power, there are economic inefficiencies with over production in non-peak periods - particularly at night. Power companies deal with that by offering large discounts for off peak only power supply. By doing so they level out the power demand. Having done so, that economic distortion then becomes the basis for insisting that renewables be 'dispatchable', but that is primarilly a demand that economic patterns geared to fossil fuel (primarilly coal) power production be preserved for power generation when it is no longer economic.
More sensibly, under primarilly renewable energy production, the discount will be shifted from the middle of the night to the middle of the day, or (with a slightly more sophisticated distribution system) to periods of peak supply. For some industrial uses (desalinization near desert regions, generation of hydrogen through electrolysis) time of supply is almost an irrelevance so that great advantage can be taken of the relatively cheap energy. For others, on-site storage will become economical with the relatively cheap electricity at peak supply.
Domestically, heating of water can obviously be done at any time. Using thermal mass from stone, so also can the heating of households, and air cooling requirements tend to coincide with excess production from renewables. The use of slow cookers can shift the peak time for cooking (the largest daily peak).
The economic inefficiencies of variable supply are not as great as those from excess supply from 'dispatchible' supply - but they are not as great as is often suggested. Certainly an industrial economy could be run on purely non-dispatchible supply. In the end the argument from the fact that many renewable sources are non-dispatchible comes down to an insistence that because we have adapted our economy to FF energy sources, we must set those adaptions in stone; and not likewise adapt to renewable sources when they become dominant.
This article leaves me very uneasy. I agree with jpjmarti @4 that it undermines the quality of the SkS site. The article title is "The strong economics of wind energy". However, the fundamentals of levelized cost are not even explained or adequately addressed. In the comments (where the issue is left to us readers) we seem to gloss over the extent to which we must over-build (x2?) to compensate for regional variations in wind/solar generation, then get drawn off into a debate about time-of-day-based dispatchability, rather than one tied to current wind speeds.
I'd love to be able to make a strong case for wind energy, and know its limits, but frankly this article and discussion lacks the depth to face the fact that wind is intermittent and any place that relies on it will enevitably find times when huge regions have insufficent wind to generate even a small fraction of the needed power. Not always, but it will occur.
There are so many issues to be addressed in any strong case for wind power: Do we need to advocate brown-outs or rotating blackouts in times of low wind? How do we need to size alternate generating capacity? How do we get the market machanisms to make best use of intermittently low supply? How well are various regions suited to wind generation, based on low correlation of wind speeds over a reasonable power tramsmission distances? The devil is in the details.
This site is called "Skeptical Science", and over the years has provided an amazing resource of information, facts, research and de-bunking. But let's remain skeptical and not suspend disbelief because something is low-carbon.
Writing blog posts from an interview with a lobby group gets a bit of a different result from writing blog posts from reviewing a number of papers.
LCOE for conventional electricity generation (i.e. dispatchable) is already a function of a few variables and you need to understand which ones were used to get your favorable / unfavorable comparison.
These variables are:
- capital cost
- interest rate, because you have to borrow money for the capital cost
- lifetime of the plant
- fuel costs
- operations and maintenance costs
It’s pretty easy to move costs around by a factor of 2-3 with the “right” choices of variables. If you want a plant with much higher capital cost to look better, choose a low interest rate or long lifetime. If you want a plant where most of the cost is fuel to look worse, choose a high fuel price. That’s easy to do as well, just find a forecast or past price you like.
If we look at the comparison between wind and gas using the conventional LCOE approach, there are 2 important factors.
First, gas is an expensive fuel, much more so than coal. The recent price difference for gas in Europe vs US is a factor of around 3.
Second, the capital cost of say a CCGT plant is quite low compared with wind so interest rates and lifetime don’t affect the CCGT LCOE very much, but the wind LCOE moves a lot.
These factors by themselves make it hard to compare the relative costs with “one number”. Better to understand each of the numbers that go into the comparison (see links below).
But then we get to the problem that LCOE is not a good metric for intermittent renewables like solar and wind.
If we take wind for example, the same turbine can be installed in Germany where the average capacity factor is under 20%, or in Oklahoma, where the average capacity factor is over 40%. Exact same turbine, similar installation costs, similar grid connection costs. But less than half the output in Germany compared with Oklahoma. This means the LCOE changes by a factor of more than 2 between these locations. Install in Ireland in a good location and you get just over 30%, and so on.
Then we have the transmission costs. Most grid operators in Europe have to pay the costs of connecting to the grid. But they don’t pay the costs of building the transmission line to get to their wind farm (not usually, e.g. Spain where the cost of building or upgrading transmission is “socialized”, the term they use). Conventional gas plants can be built near the load center, but the best location for wind farms is often far from load centers.
And transmission lines are very expensive.
Then we have the costs of intermittency (backup by plants that need to ramp up and down to cater for low/high wind periods). These are low when wind penetration is low - but obviously the plan is to get to a high penetration. These costs increase with penetration. (And there are other factors like fault ride through - becoming more common on current generation of wind turbines; frequency control in a synchronous network not yet really resolved for high wind penetration..)
This means that the real cost of wind **depends** on the case in hand (capacity factor of wind turbines for that location, current penetration of wind onto the network, costs of transmission lines that need to be built).
All in all, LCOE is not a good metric for comparing wind and gas plants unless you want to promote your point of view and then it’s very handy.
Not a simple subject. I’ve written a few articles on this subject, most relevant are:
Renewables IX – Onshore Wind Costs
Renewables XI – Cost of Gas Plants vs Wind Farms
And also an interesting example in one location, where Budischak et al (2013) calculated the lowest cost of high penetration of renewables and needed to overbuild a lot (2-3x more wind turbines). If you read the paper in detail and look at the different scenarios you can see that changes in different costs result in huge changes in the optimized build, and big changes in the total cost. None of this is captured with LCOE because it was a metric designed for a different world.
stevecarsonr... I'm curious about your Oklahoma/Germany comparison.
I know with aircraft (which includes turbine style engines), all maintenence and useful life are measured based on time used. Does the same not apply to wind turbines so that a wind turbine that gets half the use would last twice as many years?
Rob,
Interesting point. If your suggestion is correct, then capacity factor is nicely balanced out by lifetime of equipment. If you find some good information in the future I will be very interested, so please post it in a comment on one of the blog articles I linked.
I don't have any special knowledge about wind turbines in the field so I don't know if there are lifetime metrics available on current wind turbines that will be useful for this. It's quite a new field. Based on knowledge of other rotating equipment I would guess that many factors as well as run time might be important - number of starts, gustiness of wind..
The comparison of Oklahoma vs Germany is not 20% vs 40% just because of hours run, although that is part of it - it's also wind speed while running.
I'm sure that big wind farm operators are building up some metrics but they probably treat it as commercially sensitive information (i.e., competitive advantage). My experience with maintenance of equipment is that there are a lot of unknowns and typically the easiest metric to measure that clearly has an impact (e.g., run time) is the one that gets used as a proxy for asset life.
So.. maybe you are right. Maybe Germany will get 2x the lifetime for their equipment compared with Oklahoma. Maybe it's 1.1x the lifetime. Maybe on average they get 4x because of tornados and more violent storms in Oklahoma..
Charlie A:
As regards to onshore wind, the most economic on the chart, isn't usually consistently windy on the beach? 39% of the US population live directly on the shoreline right by the grid.
Thanks Steve... I'll poke around and see if I can find any information on that. Since posting that comment, I was also thinking the same might apply to the blades as well. Aircraft airframe hours are tracked, and in recent decades a lot of composites have been used in aircraft airframes. I remember in the early days of composites no one was sure what the reasonable lifetime was on the materials, but I believe that's changed with experience.
The OP mentions small wind set ups. I lived on a sailboat with both solar and a wind generator and I found that the solar was much easier to use. When the wind turbine generated power it often generated a lot of power, but sometimes we had no wind for long periods.The solar output was limited by the number of panels, but it was consistent every day. It was easy to adjust the power used to the number of panels. If we needed more power we bought another panel. The smallest solar is also much cheaper than small wind generators so it works well in undeveloped areas where people have little capitol.
The most important issue with the wind turbine was that when there is enough wind to generate a significant amount of power there is a lot of wind. At sea we almost always had a large excess of wind energy. Sailboats at anchor seek out locations that are protected from the wind. Therefore at anchor there is usually not much wind (we spent most of our time anchored). How many houses do you see built on the windy sides of ridges? More houses are built in sheltered areas. The areas where wind is best for generation are usually not very good for building houses because they are too windy. Utility setups also cover each other by setting up over large areas, as described in the OP. Small setups cannot cover each other that way, since they are small.
In addition to the caveats mentioned by stevecarsonr, there are also these factors to consider:
1. The source of LCOE presented here, Bloomberg New Energy Finance, is in the business of selling stock in renewable energy companies. In other words, they have an agenda. There are many, many sources for LCOE calculations, so one must wonder why this particular one was chosen. For example, EIA, a branch of the US DOE, does their own annual computations and gets quite different results. But even better is to not use any single analysis at all, but use a combined result of many different analyses, such as the OpenEI Transparent Cost Database, which includes Bloomberg, EIA, and many other sources as well.
2. LCOE in general can be a deceptive metric because it is designed for investors, not for policy makers. In particular, LCOE assumes a single lifetime for loans which is generally the same (and generally 30 years) across all technologies. This is problematic in two ways: first, solar and wind generators don't last that long. The National Renewable Energy Laboratory assumes 20 years for each, and actual data from Denmark confirms a mean lifetime of wind turbines of 22 years. Therefore it is unlikely banks would give 30 year loans to these projects. Second, and more importantly, using a single lifetime for all generators unfairly handicaps long-lived generating technologies (primarily hydro and nuclear) and unfairly benefits short-lived technologies (wind and solar). A hydroelectric dam can last a century or more, and most of that power is produced long after the loan is retired. That's a huge benefit to society (and to the climate) that is not reflected in LCOE.
3. "Systems costs" for renewables, such as extra grid connection costs, and extra load-balancing costs, are generally excluded from LCOE. However, the OECD has computed these costs for various technologies and their results can be found here.
4. One other factor not mentioned is that for intermittant renewables, costs will rise significantly as market penetration increases beyond the "curtailment point", which is the point at which renewables are capable of generating the entire system load. (The curtailment point is roughly equal to the capacity factor of the technology). Once renewables are built beyond that point, their capacity factor must decrease, which increases price. Since capacity factors for wind and solar are low, expect that point to be reached fairly early in the process.
Even before getting to that point, however, economic factors may severely limit deployment of wind and solar. That's because they are weather-dependent, meaning that when weather is favorable the market will be saturated with oversupply, driving wholesale electricity prices down. Consumers paying retail will not feel that, but producers certainly will: it's hard to make money if you're always selling your product at rock-bottom prices, and these conditions may make investors leery of investing in solar and wind as market penetration increases.
There's another way to look at this: To avert a climate catastrophe we just have to eliminate the burning of fossil fuels — whether we like it or not, and regardless of the cost. Arguing about the cost is therefore irrelevant.
Sooner or later, assuming we want to avert the catastrophe, we have to end up with a mix of renewable sources of electricity, some nuclear power, some electric vehicles, and synthetic fuels for other transport. Is that or is that not our ultimate goal?
Digby,
Cost is still important no matter how you slice it. In the first place, we will be able to build more non-fossil energy, faster, using the lowest-cost alternatives. In the second place, energy provides leverage that makes all other economic activities possible. Expensive energy therefore is a drag on the economy.
KeithPicketing @16. You state "first, solar and wind generators don't last that long." Your figures for wind are completely credible. When I researched PV whilst installling my 3kWp system in 2008 it was very difficult to get definitive figures on PV lifetimes. I would be very interested in seeing some estimates. I also have an evacuated tube solar thermal hot water system and again lifetime figures would be very interesting. Can you point me to any decent references?
Keith
What I meant is that instead of comparing renewables with fossil fuel, just introduce the renewable system that is best or cheapest or both for the relevant circumstances. Which system is best will vary from region to region. But ignore the economics of fossil fuel. In any case the latter should be taxed out of contention. Is that better?
Digby (#17),
The article that was written was "The Strong Economics of Wind Power" not "The Necessity of Installing Wind Power and Other Renewables Regardless of the Cost". I think this is what people, including me, have been addressing.
And, as Keith points out, knowing the cost is important anyway. Being ignorant about the cost will lead to more expensive decarbonization.
In #20 you say "But ignore the economics of fossil fuel. In any case the latter should be taxed out of contention."
Do you want to make it infinitely expensive? This will also lead to very bad outcomes. In Budischak et al (2013) you can see the (very high cost) problem of moving to 99.9% renewables. And to move to 100% from 99.9% perhaps the cost will be yet more amazing (in a bad way).
I believe being realistic about cost and technical problems is important. I seem to be in a small minority. This observation is based on anecdotal evidence only, no scientific study, but I observe:
a) article written about how cost of renewable electricity generation is actually less than conventional electricity
b) some people, occasionally me, point out that the cost calculation is problematic
c) other people arrive and say, but - item b) people - that's irrelevant: without making this change it will be the end of days
All unsurprising, but my key point is that prior to item b) the item c) folk don't arrive and say "that's irrelevant - i.e., your article about wind/solar/etc energy being lower cost - without making this change it will be the end of days". Perhaps a research project for some psychology PhDs.
villabolo (#13):
It's true that wind is "more consistently windy" near the coast as a very very general rule (except Oklahoma has the best onshore wind stats), which is why offshore wind would be wonderful if it wasn't for the high cost of offshore installations.
But - important point in understanding wind energy - if you look at a time series over a decade you usually find one period of many days of very low wind.
This is true even if you consider a very large area where the correlation between sites is apparently very low.
In Renewables XII – Windpower as Baseload and SuperGrids I had a look at a paper which investigated a supergrid around Europe/N. Africa. Then in the comments someone highlighted Kempton et al 2010 - a better paper for that subject - and I posted some data from that paper. It looks at the stats from a (hypothetical) offshore wind farm running from Maine to the Florida Keys and even with this "supergrid" the wind power production runs into a bad week over a 4 year period.
This isn't a tragedy for wind. It just indicates that you need to review time-series data rather than just rely on hand-waving, or even aggregated correlation data. The upshot for a working grid - where electricity demand is met rather than blackouts - is that either you need very expensive storage or fossil fuel backup (the latter is currently more realistic).
Steve
Okay, it looks like I should refine my viewpoint. I admit I was looking at the end goal in the future. Some people say we've got 20 years to reduce fossil-fuel use to near zero; others extend the period, but it's still relatively short. Whatever the period, the taxing of fossil fuel needs to increase steadily so as to effect this result. I don't mean that it should happen overnight. Is that correct?
Digby (#23),
How to reduce fossil fuel usage - now there is an economic argument.
Putting a price on carbon is endorsed by many economists as simpler and more efficient than other approaches (like cap and trade). But I can't claim to any real understanding of economics despite a lot of effort in that direction on my part - uncertainties seem to overwhelm calculations and falsifiable theories.
My only point is that realistic costs surely help in any assessment, rather than just adding an infinite tax.
How quickly should the world reduce CO2 emissions to zero? No idea. Add that uncertainty to the huge uncertainty that is economics and I think (but can't prove) that you can get any answer you want.
Steve Carson,
I looked over your post at Science of Doom and noticed that your primary references are out of date. Czisch & Ernst (2001) was the one you showed the most data from. Did they use hub heights of 90-100 meters, like current wind turbines use, or did they use 50 meters as was the case in 2001? Is this data really applicable to the current state of the art? I doubt it.
In Kepton (2010) they rely on wind from a number of locations in Florida. Nowhere in Florida has good enough wind to be commercial so that data appears to expand their areal coverage without adding any sites where someone who wants to generate electricity would build a wind farm.
On the other hand, Budichak (2013), a more up to date paper, whch you cite above (why isn't Budichak mentioned at Science of Doom?), uses data to show that over a 4 year period they are able to use wind and solar for up to 99% of electricity. While Budichak is more expensive to build out all necessary power, they explicitly state that they do not use hydropower from Canada as a backup because it would be too easy to use renewables for 100% of power in that case. They also do not connect to other nearby grids which would also make it easier and much cheaper to build out the overall grid. They wanted to demonstrate that renewables could be used, not to find the cheapest possible grid. Obviously it is more economic to build the cheapest grid.
Since you neglect to mention Jacobson et al (2014) which is more the state of the art I am not sure what you find objectionable with that analysis. Jacobson and his group have evaluated the entire USA using hourly data, as you have said is necessary, and found that it is cheaper to use 100% renewable power. They use a great deal of Hydrogen in their plan, which I think is risky, but many other opitons are available. For example, excess electricity could be used to make diesel fuel instead of hydrogen.
It seeems to me that a realistic evaluation of renewables should include current research and not rely primarily on references from 2001 and a source that relies on Florida wind for a significant part of its power. 2001 cannot be state of the art and Florida is not economic for wind.
I am disappointed that Science of Doom chose to use outdated references for their analysis. A brief check of the papers that cite Budichak would give many references that are much more up to date.
When considering the economics of the size of a wind turbine I suggest considering the basic equation that determines the energy capture of a blade size:
Power = 0.5 x Swept Area x Air Density x Velocity3
Swept area is the most important factor when it comes to cost.
The bigger the turbine the better, because if you double the blade length the swept area is the square of that length (Pi x r2).
With wind turbines economies of scale are also supported by maths/physics.
A game changer though is the field of superconductors. A superconducting wind turbine of say 3MW would probably be half the weight/size of todays turbines. It's a shame American Superconductor ran into problems a few years ago (with dodgy Chinese license agreements), otherwise we may already have a working superconducting wind turbine.
I was noticing in the Seimen's LCA docs I was reading through that, end-of-life for wind turbines is based on timing of newer technology. Older turbines are taken down, refurbished and sent to new locations, and the existing site will have a newer, larger turbine installed.
That seems to build a whole other factor into the cost of scaling wind power.
You mean, in a good way?
StBarnabas (#19)
System lifetimes for solar are notoriously hard to pin down, and may vary considerably by both manufacturer and by the skill of the installers. One old but useful reference is Czandernda & Jorgensen (1997), who point out that UV degradation of the cell is only one of several possible failure modes. For real-world data from a utility-scale installation, try Mallineni 2013. YMMV.
Digby (#20)
Point taken, as long as we substitute "non-fossil" for "renewable". I'm not leaving out nuclear, and I hope nobody else does either.
Michael Sweet (#25)
The problem with Mark Z. Jacobson is that he won't tell us how much his all-renewables plan costs in numbers, prefering only adjectives instead. The adjective is invariably "low", a characterization he arrives at by applying the external cost of fossil fuel. That's not necessarily wrong, but it certainly is inadequate. Budischak is more open, telling us that all-renewable will roughly triple the price of electricity. Based on the recent experience in Germany and Denmark, I can certainly believe it.
Meanwhile, the Deep Decarbonization Pathways Project (which is interested in decarbonization, and not just all-renewable-at-any-price) finds that the median high-renewable pathway is about four times more expensive than the median high-nuclear pathway. Not surprising, when you consider that high-renewables must depend on either storage (very expensive) or transmission (very expensive) to cover the windless nights. Their high nuclear pathway also sees a large increase in renewables from present levels, but keeps it below the curtailment point, and therefore keeps systems costs at a minimum.
So the lowest cost alternative appears to be (1) hydro and geothermal where available; (2) wind & solar up to the curtailment point; and (3) nuclear for the rest. Also bear in mind that as we decarbonize the grid, we will decarbonize currently non-electric fossil use by switching to grid electricity (e.g., cars and space heat). Thus we should expect electricity demand to increase significantly as we decarbonize.
Regarding efficiency, it has never reduced overall energy demand in human history, and recent works by Garrett suggest it never will. Thermodynamics can be a dismal science.
bozzza... Can you reference which comment you're responding to?
Rob Honeycutt (#27)
You should be aware that a larger wind turbine does not necessarily mean cheaper. Power available to a wind turbine scales with the square of the rotor diameter, but cost of the turbine scales with mass, which in turn scales with the cube of the rotor diameter. There is some improvement by getting big enough to get the turbine out of the surface boundary layer, but that improvement drops off rapidly with height.
The upshot is that there is a "sweet spot" for turbine size that gets you the lowest price per kWh generated. The sweet spot varies by capacity factor. For most locations, that sweet spot is in the 1.5 to 3 MW range. Offshore turbines can be bigger, but that's also reflected in the higher cost of offshore wind power.
Keithpickering... California has most certainly used efficiency to reduce per capita electricity consumption. LINK
keithpickering @31...
That brings up a question I'm trying to answer right now. Currently the stated lifespan of a wind turbine is 20-25 years. But there has to be more to it than that. What are the factors that determine that lifespan? If what you say is correct, and turbines have reached their "sweet spot" then what's to stop them from just refurbishing in place? Surely the tower itself can be constructed to last in excess of 100 years. The composite blades I would expect have a useful lifespan in excess of 25 years (or 100,000 hours). The gear box and generator could easily be overhauled (or upgraded) and run for another 25 years.
michael sweet (#25)
1. Why Czisch & Ernst (2001)
I tried hard to find papers which looked at huge grids.
The reason for this is even with country-wide grids in Europe there would be multiple days where electricity production would be low, even if the whole grid was tied together with enough transmission capacity to allow one region of the country to power the entire demand of the whole country.
To find other papers with better results I looked at many that cited Czisch & Ernst (2001). This was mainly because Czisch & Ernst didn't quite answer the real question I wanted answered, although they gave indications of the problem.
Relating to your claim/question about wind turbine heights, here is what they say in their paper: "For this study the two of these levels close to 33m and 144m above ground were used to calculate the world-wide wind conditions at 80m hub height. The wind data were converted to power using the characteristics of a wind turbine (WT) with variable speed, 80m hub height 1.5 MW capacity and 66m rotor diameter."
So this indicates they are not quite looking at 90-100m, but they are not considering 50m. In any case, the question is not about the total power possible from a grid, it is about how to get something more like baseload power. How much does the height moving from 80m to 95m affect multiple low wind days? I have not seen anyone address this question.
2. Kempton et al 2010
I was very pleased to have a commenter bring up this paper in the discussion and so cited from it extensively in the comments on the Czich paper. You say "..In Kepton (2010) they rely on wind from a number of locations in Florida.."
Perhaps you are looking at a different paper. The paper I cited in the comments of my article and showed a map is called "Electric power from offshore wind via synoptic-scale interconnection" by Willett Kempton, Felipe M. Pimenta, Dana E. Veron, and Brian A. Colle and published by PNAS.
This paper considers a grid from Maine down to the Florida Keys. Basically the entire offshore east coast of the USA. Maybe you want to revise your comment.
3. Budichak (2013)
You ask "..On the other hand, Budichak (2013), a more up to date paper, whch you cite above (why isn't Budichak mentioned at Science of Doom?)"
I did a whole article on Budichak (2013): Renewables XIV – Minimized Cost of 99.9% Renewable Study
Why isn't Budichak mentioned at Science of Doom? Maybe you want to revise this question as well.
Then you say:
"..They also do not connect to other nearby grids which would also make it easier and much cheaper to build out the overall grid. They wanted to demonstrate that renewables could be used, not to find the cheapest possible grid.."
The first part of that statement is an untested claim (and the second part is not really correct). As you can see in the article I wrote on Budichak's paper, I question this. They do not attempt to prove it in the paper.
I did speak to Cory Budichak after writing the paper as I had emailed him and asked for his comment on a number of points including that one. He suggested that larger grids didn't have low wind days due to low correlation and pointed to Kempton et al 2010. But Kempton really demonstrates the opposite.
Actually their problem was a little different, as you can see if you read the paper in detail. The computing resources constrained the optimization. Therefore, the cost of transmission could not be included in the calculation. They already had something like 2 billion scenarios to calculate.
It would be wonderful to see someone redo the calculations for Budichak's scenario including a transmission cost for a larger grid. Does more transmission cost and better wind production reduce overall cost or not?
I have yet to find such a paper, perhaps one has not been done. It isn't trivial. Transmission lines are very expensive and given the data calculated by Kempton et al 2010 it looks like even a much larger grid than the PJM network will still have many consecutive days of very low wind production. And cost a lot more.
If the solution is simple - just build a larger grid and that solves the problem - then this should be easy to demonstrate.
Budichak's paper is worth studying in detail. I recommend it.
Steve, Keith
Upon further reflection I return to my original position — with a modification. In a rational world it really is irrelevant comparing the cost of non-fossil energy with fossil fuel. In such a rational world we would phase-in non-fossil energy and phase out fossil fuel, regardless of cost, because it is the rational thing to do, because logic dictates that doing so would avoid enormous costs in the future.
The analogy that comes to mind is that of slavery. I don't know if it's a valid analogy, but here goes:
We could use slave labour to power the economy because it is cheap, but we do not because that is the moral thing to do.
Similarly, we could continue to use fossil fuel to power the economy because it is cheap, but we should not because that is the rational thing to do.
Unfortunately we do not live in a rational world. We are trapped in a culture ruled by the dollar. So that's why we have to resort to monetary trickery to persuade ourselves to do the right thing. Somehow I don't think it'll work.
Keith Pickering at 29:
I do not see the description of the numbers in the Deep Decarbonization Pathways Project that you claim. It seems to me that you have misread figure 12 on Incremental Energy Costs.
On page 47 the estimated yearly costs for the high nuclear scenario is $20 billion for nuclear and $30 billion for renewables. For the renewable scenario the estimate is $70 billion for renewables alone. The cost of renewables is about 40% higher, not 400%. I will note that they estimate the cost to build about 300 nuclear plants in 30 years, 10 plants per year, at a cost of $20 billion per year. That is $2 billion per plant. The Vogtle Plants, current state of the art, are estimated to cost $3.5-4.0 billion each. I will leave it up to other readers to decide if the estimated cost of nuclear plants is reliable and if the nuclear industry can build 40-60 plants at the same time when they currently struggle to build 4 plants. I cannot evaluate their claims for cost of renewables.
The best news from this analysis is their conclusion that all of the scenarios that they evaluated could potentially lead to the desired cuts in carbon emissions by the year 2050. This means that if we get serious and begin to build out a carbon free system there are several ways that can succeed. If we start out to build renewables and it is too expensive we can switch to nuclear. If nuclear is too expensive we can build more renewables. If carbon capture is economic it can work. It will certainly be more clear in five years what is the best path. Now we should build whatever we can as fast as we can.
The idea that we don't live in a rational world is not always correct. Money has to flow to create jobs and profits- for those creating those jobs. If employers aren't incentivised with the fruits of enterprise then they just stash their old money away never to see the light of day.
For Michael and others, on nuclear:
There's a very good summary of nuclear in an open acccess paper, Joskow, Paul L., and John E. Parsons. “The economic future of nuclear power.” Daedalus 138.4 (2009): 45-59. http://dx.doi.org/10.1162/daed.2009.138.4.45
And also a good summary in a very recent paper (not journal paper) published by JP Morgan for investors in renewable energy:
Our annual energy paper: the deep de-carbonization of electricity grids
This paper does a great job of capturing many of the renewable electricity issues and is well worth reading. They comment on nuclear:
And then some more detail in their Appendix V, with the final comment:
Digby @35... I would suggest that people sometimes forget that there are already significant social costs attached to our use of fossil fuels. Some estimates range as high at $200/ton. By that measure, carbon-free energy solutions are far cheaper than fossil fuel sources. We just current lack the methods of attaching those costs to their sources.
By implementing a carbon tax and ramping it up over time, we are potentially making economies more efficient. If those estimates are anywhere near accurate, then we needn't be concerned so much about the cost. The biggest challenge is the upfront costs of transition.
Rob Honeycutt #33
Total energy use and per-capita use are different metrics.
Rob Honeycutt #34
Wind turbine components are subject to enormous physical loads, and those loads are cyclical. This is just hell on any material. Refurbishment frequently is not an option when facing metal fatigue, you've got to replace. High-speed, low-torque generators are subject to much less wear and tear. Low-speed, high-torque is the worst. Combine that with cyclical loading and it's the worst of the worst. This is a case where smaller is better, because smaller generators have higher RPM and lower torque.
michael sweet #36
No, actually you're misreading. The numbers reported on page 47 are annual capital costs, which are just one component of the total system costs reported on Figure 12.
keith @40... You're describing exactly what aircraft already do. Typical lifespan for a commercial airliner airframe is 120,000 hours. Composite airframes can apparently go even longer, I believe, but complete composite airframes haven't been used in commercial aircraft yet. And, no, the loads shouldn't be that much different than commercial aircraft. In fact, aircraft likely have greater load factors to deal with than wind turbines, and are designed to take on extreme loads due to the fact they carry passengers.
keith... I'd also suggest the torque factor is controlable through blade pitch. Torque loads are going to be a function of the resistence presented by the turbine the blades are pushing relative to how hard you make the blades push it. Right?
keith... RE: per capita and total energy. If California's population had stayed flat through that period then total energy use would have fallen.
Rob #41, 42
Stress on an airframe is very much smaller than stress on wind turbine components, and much less cyclical. Reducing torque on a turbine is possible, but that loses energy output, which is the whole point of the machine. (Power = torque x RPM).
#43
Doubtful. Total energy consumption is much more closely tied to GDP than to population. Of course, we could reduce (and have reduced) total energy consumption in periods of economic contraction, but that's hardly an acceptable solution. And in fact it's not a solution at all, since any climate plans that require economic contraction would be unable to garner enough political support for their implementation. This again points up the need to keep cost as a touchone when considering alternatives to fossil fuel.
Keith... How do you make the determination that the stresses are less? I would expect a typical commercial aircraft is designed to take far more stress and cycles than a wind turbine. Passenger aircraft are designed to take in excess of 6 G's fully loaded. They're designed for high negative G-loads, and are clearly designed for cyclical stress loads for turbulent air and for repeated takeoff and landings. The cyclical stresses on a wind turbine blade can't be nearly as high, and is likely more limited to the differences between the force placed on the high and low blades due to laminar flow.
As I originally mentioned, according to Seimen's LCAs, the 20-25 lifespan is based on a tear down, refurbishment and transport to a new location, while the original location would be replaced with updated equipment.
[Edit]
Note that the force exerted on an airframe squares with speed. Large commercial passenger aircraft have a max takeoff weight of upward of half to a million pounds and cruise at speeds of 500 knots. That's a massive amount of force when you enter unexpected turbulence.
Rob #39
Yes, we are now struggling with the idea of attaching the costs of pollution to the sources, but we have not yet succeeded. This comes of a culture ruled by the dollar and one that views the planet as an infinite repository of resources and a bottomless sink for pollution. I suspect we'll need to change the culture in order to make any further progress. Continuing climate change might not give us the necessary time.
Digby... I would disagree. I think we're likely very close to seeing broad use of carbon taxes.
Rob, to date we have not succeeded, but I hope you're right and things will start to happen soon. I'll wait and see.
keithpickering (#29) many thanks for the PV lifetime references. Sadly the one with actual data, Mallineni (2013) is for Arizona (which I know well from my PhD days) is for a very different climate to Northumberland; a lot colder and wetter. I will keep my eye on the literature.